for - explanation - why minimising surprise is a good definition of intelligence - Donald Hoffman - it's very intelligent to minimize surprise - I'm surprised all the time - I'm pretty stupid right, I don't understand the world very well - but if I'm NOT surprised, it's like I've got a really good model especially if I'm doing lots of stuff in the world and I'm almost never surprised - boy am I I'm really intelligent! - So, you can see why that's a really good principle for trying to build an AI, - not just finding correlations between everything, - but really something deeper.